#11484: Type synonym using -XTypeInType can't be spliced with TH -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Template Haskell | Version: 8.0.1-rc1 Resolution: | Keywords: TypeInType Operating System: Unknown/Multiple | Architecture: Type of failure: GHC rejects | Unknown/Multiple valid program | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by goldfire): My guess is that this problem is a consequence of the quite delicate design of `TcHsType.splitTelescopeTvs`. I imagine there's a way to apply a delicate patch to the delicate algorithm to fix this, but I propose a better solution that makes this all more robust: refactor `TcTyCon` to store more information about user-written telescopes (that is, lists of type variables that may have dependencies) so that `splitTelescopeTvs` becomes trivial. When I first wrote `splitTelescopeTvs` as part of the `TypeInType` work, I hadn't yet added `TcTyCon`, and I was very loathe to put surface-Haskell details about a user-written telescope into `TyCon`. But now that we have `TcTyCon`s that exist only in the first stages of type-checking, it's more reasonable to put in more surface-Haskell information. I think this refactoring would be fairly straightforward, but I'm afraid I haven't the time to do it myself. Happy to advise if you (for any value of "you") wish to take this on. Alternatively, if you want to figure out how to fix this particular problem without doing the larger refactor, I'm happy to look at that, too. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/11484#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler